Don't search titles with regexes.

[utils.git] / ANN.pm

diff --git a/ANN.pm b/ANN.pm
index 2ae48cd..1406c45 100644 (file)
--- a/ANN.pm
+++ b/ANN.pm
@@ -1,21 +1,45 @@
 package Anime::ANN;
 
+use POSIX;
+use Digest::MD5;
 use LWP::UserAgent;
 use HTML::Entities;
 
 $ver = "0.1";
+$ua = LWP::UserAgent->new;
+$ua->agent("ANNData/$ver ");
+$usecache = 1;
 
 sub _get
 {
-    my($uri, $ua, $res);
+    my($uri, $cname, $res);
     ($uri) = @_;
 
-    $ua = LWP::UserAgent->new;
-    $ua->agent("ANNData/$ver ");
-
+    $cname = $ENV{"HOME"} . "/.ann/";
+    mkdir $cname unless -e $cname;
+    $cname .= "cache/";
+    mkdir $cname unless -e $cname;
+    $cname .= Digest::MD5::md5_hex $uri;
+    if($usecache && -e $cname) {
+       my(@s);
+       @s = stat $cname;
+       if((time - $s[9]) < 86400) {
+           my($buf);
+           open CACHE, "<:utf8", $cname;
+           $res .= $buf while read CACHE, $buf, 1024;
+           close CACHE;
+           return $res;
+       }
+    }
+    
     $res = $ua->request(HTTP::Request->new("GET", "$uri"));
-
-    die "could not fetch $uri\n" unless $res->is_success;
+    
+    if(open CACHE, ">:utf8", $cname) {
+       print CACHE $res->content;
+       close CACHE;
+    }
+    
+    return undef unless $res->is_success;
     return $res->content;
 }
 
@@ -26,14 +50,21 @@ sub getlist
     
     $il = uc(($name =~ /^(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
-    $html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il";
+    if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
+       return undef;
+    }
     
     # The only way to recognize entries that seems sure is to look
     # after the "HOVERLINE" class.
     
-    push @ret, $1 while $html =~ /<A\s.*CLASS=HOVERLINE\s.*>.*<FONT.*>([^<>]*$name[^<>]*)<\/FONT/ig;
+    while($html =~ /<A\s.*CLASS=HOVERLINE\s.*HREF=\"([^\"]+)\".*>([^<]+)<\//ig) {
+       if((substr "" . lc $2 , 0, length $name) eq lc $name) {
+           push @ret, $2;
+       }
+    }
+    # push @ret, $1 while $html =~ /<A\s.*CLASS=HOVERLINE\s.*>.*<FONT.*>([^<>]*$name[^<>]*)<\/FONT/ig;
     
-    return(@ret);
+    return @ret;
 }
 
 sub getid
@@ -43,14 +74,20 @@ sub getid
     
     $il = uc(($name =~ /^(.)/)[0]);
     $il = "9" if (!($il =~ /[A-Z]/));
-    $html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il";
+    if(!($html = _get "http://www.animenewsnetwork.com/encyclopedia/anime.php?list=$il")) {
+       return undef;
+    }
     
     # The only way to recognize entries that seems sure is to look
     # after the "HOVERLINE" class.
     
-    (($url) = ($html =~ /<A\s.*CLASS=HOVERLINE\s.*HREF=\"([^\"]+)\".*$name/i)) || return;
+    while($html =~ /<A\s.*CLASS=HOVERLINE\s.*HREF=\"([^\"]+)\".*>([^<]+)<\//ig) {
+       if((substr "" . lc $2 , 0, length $name) eq lc $name) {
+           return ($1 =~ /id=(\d+)$/)[0];
+       }
+    }
     
-    return ($url =~ /\?id=(\d+)$/)[0];
+    return undef;
 }
 
 sub geturl
@@ -70,8 +107,8 @@ sub getthemes
        my(@parts, $ct, $buf);
        while($html =~ /\G\<br\>&nbsp; &nbsp; (([^<>]|\<i\>|<\/i>)+)/igc) {
            $buf = $1;
-           #                     0  1            2         3     4     5        6        7                     8              9         10     11
-           if(@parts = ($buf =~ /(\#(\d+):)?\s*\"([^\"\(]+)(\s+\((\<i\>(.*)<\/i>(;\s*)?)?([^<>]+)?\))?\"\s+by\s+([^\(]*[^\(\s])(\s*\(eps (\d+)-(\d+)?\))?/i)) {
+           #                     0  1            2           3     4     5        6        7                      8              9         10    11
+           if(@parts = ($buf =~ /(\#(\d+):)?\s*\"([^\"\(]+\S)(\s*\((\<i\>(.*)<\/i>(;\s*)?)?([^<>]+)?\))?\"\s+by\s+([^\(]*[^\(\s])(\s*\(eps (\d+)-(\d+)?\))?/i)) {
                $ct = {};
                $ct->{"num"} = $parts[1] if defined $parts[1];
                if(defined $parts[5]) {
@@ -97,7 +134,9 @@ sub getseries
     my($id, $buf, $html, %ret);
     ($id) = @_;
     
-    $html = _get geturl $id;
+    if(!($html = _get geturl $id)) {
+       return undef;
+    }
     
     $ret{"url"} = geturl $id;
     ($buf) = ($html =~ /\<TITLE\>Anime News Network - ([^<]*)<\/TITLE>/);
@@ -112,6 +151,9 @@ sub getseries
     if(($buf) = ($html =~ /number of episodes:<\/b>\n([^<]+)</is)) {
        $ret{"eps"} = $buf;
     }
+    if(($buf) = ($html =~ /genres:<\/b>\n([^<]+)</is)) {
+       $ret{"gnr"} = [split /, /, $buf];
+    }
     $buf = getthemes $html, "opening";
     $ret{"op"} = $buf if(@{$buf});
     $buf = getthemes $html, "ending";